[BioC] edgeR and pair-wise comparison error by estimateGLMCommonDisp

Sheng [guest] guest at bioconductor.org
Mon Nov 7 01:07:17 CET 2011

Hi All,

I am trying to use edgeR to find differentially expressed genes from a pairwise comparison. So I am following the edgeR's manual chapter 13: Case study: Oral carcinomas vs matched normal tissue. However, when I am trying to "estimateCommonDisp" to estimate the Cox-Reid common dispersion, I got an error:
"Error in beta[k, ] <- betaj[decr, ] : 
  NAs are not allowed in subscripted assignments" 

Does anyone have ideas about this? I have 6 paired data. Thanks a lot!

My code is following:

x = X[,-(1:2)]
geneid = X[,1]

group=factor(rep(c("N", "C"), 6))

# build edgeR object
d = DGEList(x,group=group, genes=rownames(x))

# filter out genes have fewer than 1 count per million
cpm = cpm(d)
d <- d[rowSums(cpm > 1) >= 3, ]

# plot MDS
# plotMDS(d, main="MDS Plot for Tuch Data")

# design matrix
patient = factor(rep(1:6, each=2))
design <- model.matrix(~patient+d$samples$group)
rownames(design) <- rownames(d$samples)
colnames(design)[7] <- "Cancer"

# calculate 
d <- estimateGLMCommonDisp(d, design)


 -- output of sessionInfo(): 

R version 2.14.0 (2011-10-31)
Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)

[1] C/en_US.UTF-8/C/C/C/C

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] limma_3.10.0 stringr_0.5  edgeR_2.4.0 

loaded via a namespace (and not attached):
[1] plyr_1.6     tools_2.14.0

Sent via the guest posting facility at bioconductor.org.

More information about the Bioconductor mailing list